Hi-Res Audio Explained: Do 24-Bit and 192 kHz Matter?

Streaming services now sell the same album at 16-bit/44.1 kHz, 24-bit/96 kHz and 24-bit/192 kHz, and the bigger numbers come with a bigger price tag or a bigger badge. The numbers are real, but what they mean for your ears is much less obvious than the marketing suggests. To judge them, you need to know what sample rate and bit depth actually measure, where each one stops mattering, and what the listening tests have found.

What sample rate and bit depth actually measure

Digital audio takes thousands of snapshots of a sound wave every second. The sample rate is how many snapshots per second. The bit depth is how finely each snapshot's loudness is recorded. Think of a photo: sample rate is how often you take a picture of a moving object, bit depth is how many shades of gray each picture can hold.

CD audio uses 44,100 samples per second and 16 bits per sample. The common "hi-res" tiers are:

Format	Sample rate	Bit depth	Uncompressed data rate (stereo)
CD quality	44.1 kHz	16-bit	1,411 kbps
Hi-res (common)	96 kHz	24-bit	4,608 kbps
Hi-res (top tier)	192 kHz	24-bit	9,216 kbps

Lossless codecs such as FLAC shrink these files by roughly a third to a half without changing a single sample, so a 24/192 file is still several times the size of its CD-quality cousin. If you want the other side of the compression story, see how lossy audio compression works for what MP3, AAC and Opus throw away.

Sample rate: what it really buys you

The sample rate sets the highest frequency you can capture. The Nyquist theorem says that a system sampling at 44.1 kHz can represent everything up to half that figure, 22.05 kHz, perfectly. Human hearing tops out around 20 kHz in young, healthy ears and drops with age, often well below 16 kHz for adults. So 44.1 kHz already covers the whole audible range with a small safety margin.

Two common misunderstandings are worth clearing up.

• "Digital audio is a staircase." Sample points do not get joined by steps. The converter reconstructs a smooth, continuous wave that matches the original exactly within the captured bandwidth. You can see this on an oscilloscope: a sine wave from a 44.1 kHz converter looks like a sine wave, not stairs.
• "More samples means more detail." Above the audible limit, extra samples describe frequencies you cannot hear. They do not add resolution to the music you can.

Higher rates do have genuine uses in the studio. Recording at 96 kHz lets plug-ins that create harmonics (distortion, saturation, pitch shifting) work with less aliasing, and it relaxes the design of the filters in converters. That is why many producers work at 88.2 or 96 kHz even when the final release is 44.1.

There is a downside, too. Content above 20 kHz can push a speaker or amplifier into intermodulation distortion, where inaudible ultrasonic tones mix and produce audible artifacts. The Xiph.Org engineer Chris "Monty" Montgomery lays out this argument, along with plenty of demonstrations, in 24/192 Music Downloads and why they make no sense. Not everyone agrees with all of his conclusions, but the physics he describes is standard.

Bit depth: dynamic range, not "detail"

Bit depth does not make quiet parts "more detailed." It sets the distance between the loudest possible sound and the noise floor. Each bit adds about 6 dB of dynamic range, so 16-bit gives roughly 96 dB and 24-bit gives roughly 144 dB.

Put those numbers in context:

• A quiet bedroom at night sits around 30 dB SPL.
• A symphony orchestra at its loudest hits roughly 100 to 110 dB SPL at a good seat.
• That gap of 70 to 80 dB is already smaller than the 96 dB a CD offers.
• The best real-world converters reach about 120 dB. The 144 dB of 24-bit is a theoretical number that no analog circuit can fill, because the thermal noise of the electronics gets in the way first.

On top of that, 16-bit audio is normally made with dither, a tiny amount of shaped noise added before rounding. Dither turns the rounding error into a steady, very quiet hiss instead of harmonic distortion on quiet passages. With dither, a 16-bit file can even carry sounds below its nominal noise floor.

Where 24-bit clearly matters is in recording and mixing. When you capture a vocal, you want to leave plenty of headroom so a surprise shout does not clip. With 24-bit you can record at a comfortable -18 dBFS average level and still have a noise floor far below anything audible. Many modern DAWs also process internally in 32-bit float, which makes accidental clipping inside the software almost impossible. Those are benefits for the people making the music, not for the people playing the final file.

What the listening tests found

The most cited study on this topic is by E. Brad Meyer and David Moran, published in the Journal of the Audio Engineering Society in 2007. They ran double-blind tests over more than a year comparing high-resolution playback with the same signal passed through a 16-bit/44.1 kHz bottleneck. Listeners included recording engineers, students and dedicated audiophiles. According to the abstract, the CD-standard loop was not detected by any subject on any system at normal to loud listening levels.

The study has fair criticisms. Some of the source discs were SACD and DVD-Audio releases of older analog recordings, which may not contain real content above the CD band, and the test equipment was not the most revealing available. A later meta-analysis by Joshua Reiss, also in the AES journal, pooled many experiments and reported a small but statistically significant ability to tell the formats apart, particularly with trained listeners. The honest summary is that any audible difference is small, hard to produce under controlled conditions, and nowhere near the leap you hear between a 128 kbps MP3 and a lossless file.

A big practical problem in casual comparisons is that the "hi-res" version is often a different master. If the hi-res release has different EQ, less compression or a different loudness, you are comparing masters, not formats. Our piece on how loudness normalization ended the loudness war explains why mastering choices change the sound far more than the container does.

What streaming services actually offer

Services have settled on a few tiers. Spotify launched Lossless in September 2025 for Premium subscribers, with a ceiling of 24-bit/44.1 kHz, as What Hi-Fi noted when comparing it with rival hi-res services. Apple Music offers Lossless up to 24-bit/48 kHz and Hi-Res Lossless up to 24-bit/192 kHz, and Amazon Music, Tidal and Qobuz also go up to 192 kHz on parts of their catalogs.

Two things worth knowing before you pay for a higher tier:

1. Your chain limits you first. Bluetooth headphones compress the signal again, and most standard Bluetooth codecs do not carry anything near 24/192. Wireless earbuds make the hi-res badge irrelevant in practice.
2. The file may not be natively hi-res. A 2005 album recorded on 16-bit equipment and upsampled to 24/96 contains nothing above its original band. Look for the original recording information where the service provides it.

How to test it yourself

You can run your own blind test in about 20 minutes, and it is more convincing than any article.

1. Choose a well-recorded track with wide dynamics and real acoustic instruments, such as a jazz trio or a string quartet.
2. Get a genuine hi-res version and make a 16-bit/44.1 kHz copy of that exact file with a converter (with dither on), so both versions share the same master.
3. Level-match them. Even a 0.5 dB difference makes the louder one sound better.
4. Use a blind tool such as an ABX plugin in a player like foobar2000, and have a friend switch between versions at random.
5. Do at least 16 trials. Chance alone gives you 50 percent, so you need roughly 12 of 16 correct before the result means anything.
6. Listen on the best wired headphones or speakers you have. If you are shopping for those, open-back versus closed-back headphones matters far more for the sound you hear.

What to spend your attention on instead

If you want better sound at home, the biggest gains come from places that cost less than a hi-res subscription. Start with lossless instead of lossy streaming, since that step removes real artifacts. Then fix your listening environment, your speaker placement or your headphone fit, and use equalization only to correct clear problems. Finally, pick the versions of albums with a dynamic, uncrushed master.

Hi-res formats are a perfectly fine thing to own if you like the idea or want future-proof archives. Just treat the numbers as a spec sheet for the studio, not a promise about your living room.

Try this tonight

Pick one album you know very well and play it three ways: your usual streaming setting, a lossless setting, and a hi-res setting if your service has one. Keep the volume identical, use wired headphones, and write down what you notice before looking at the screen. If you only hear a difference when you know which version is playing, you have learned something useful about your own ears, and you can save the money for better headphones or more music.